Mathematically Rigorous Results In Sequence Matching.

点击翻译按钮获取中文摘要

The BLAST group has now incorporated our faster calculations for the Gumbel scale parameter, λ, pre-factor k, and finite-size correction into the BLAST code. The new finite-size correction demonstrably improves the retrieval order. In practice, e.g., biologists notice and find it irritating when an exact match to their query is not the highest-ranked hit in a sequence database. Our finite-size correction places identical matches more consistently at the top of the retrieval list than the old finite-size correction. Our methods are being used to compute the statistical parameters for several new DNA scoring schemes at NCBI, and we are collaborating with Dr. Martin Frith in extending our methods to next-generation sequence matching, including frameshifts in DNA.